Research & Projects

Proposed new method for online estimation of state of charge for dynamically discharged Li-ion batteries using deep neural networks. Achieved MSE in order of 10−5 between true and estimated SOC. 

Research paper on the proposed methodology published at IEEE- Power and Energy Society General Meeting 2023 conference: Online State of Charge Estimation Framework using Hybrid Equivalent Circuit Model and Neural Network. 

Developed MATLAB algorithms for states estimation in BMS for 1.5kW battery packs.

Implemented probabilistic inferential approaches (Kalman filter, Extended Kalman filter, and Sigma Point Kalman filter) for state of charge estimation of Lithium (Li)-ion battery pack.

Implemented weighted total least square, Support Vector Regression and Gaussian Process Mixtures for state of health and remaining useful life estimation of Li-ion battery pack. 

Developing a robust machine learning algorithm that leverages the In-Context Learning capabilities of large language models (LLMs) for the cybersecurity of substations. 

Utilizing weak learning to train the LLMs for creating a robust and generalizable ML intrusion detection system capable of detecting out-of-distribution (unseen) attacks.